[PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[Laurent Pinchart]

Renesas ARM platforms are transitioning from single-platform to
multi-platform kernels using the new ARCH_SHMOBILE_MULTI. Make the
driver available on all ARM platforms to enable it on both ARCH_SHMOBILE
and ARCH_SHMOBILE_MULTI and increase build testing coverage.

Don't enable COMPILE_TEST support as the driver doesn't compile on
non-ARM platforms due to usage of the ARM DMA IOMMU API.

Cc: Joerg Roedel <joro@8bytes.org>
Cc: iommu@lists.linux-foundation.org
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
---
 drivers/iommu/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
index c880eba..3e7fdbb 100644
--- a/drivers/iommu/Kconfig
+++ b/drivers/iommu/Kconfig
@@ -206,7 +206,7 @@ config SHMOBILE_IPMMU_TLB
 config SHMOBILE_IOMMU
 	bool "IOMMU for Renesas IPMMU/IPMMUI"
 	default n
-	depends on (ARM && ARCH_SHMOBILE)
+	depends on ARM
 	select IOMMU_API
 	select ARM_DMA_USE_IOMMU
 	select SHMOBILE_IPMMU
-- 
1.8.1.5

Re: [PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[Russell King - ARM Linux]

On Wed, Oct 30, 2013 at 12:20:43PM +0100, Laurent Pinchart wrote:
> Renesas ARM platforms are transitioning from single-platform to
> multi-platform kernels using the new ARCH_SHMOBILE_MULTI. Make the
> driver available on all ARM platforms to enable it on both ARCH_SHMOBILE
> and ARCH_SHMOBILE_MULTI and increase build testing coverage.
> 
> Don't enable COMPILE_TEST support as the driver doesn't compile on
> non-ARM platforms due to usage of the ARM DMA IOMMU API.

For similar reasons as x86, can we please think about using:

	depends on ARM
	depends on ARCH_SHMOBILE || ARCH_SHMOBILE_MULTI || COMPILE_TEST

This way we don't end up polluting the configuration for non-shmobile
platforms.  Same goes for other ARM stuff... the number of options is
getting rather large and we need to think about keeping that in check
where its easily possible to do so.

Re: [PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[Laurent Pinchart]

Hi Russell,

[CC'ing Mark Brown]

On Wednesday 30 October 2013 11:26:25 Russell King - ARM Linux wrote:
> On Wed, Oct 30, 2013 at 12:20:43PM +0100, Laurent Pinchart wrote:
> > Renesas ARM platforms are transitioning from single-platform to
> > multi-platform kernels using the new ARCH_SHMOBILE_MULTI. Make the
> > driver available on all ARM platforms to enable it on both ARCH_SHMOBILE
> > and ARCH_SHMOBILE_MULTI and increase build testing coverage.
> > 
> > Don't enable COMPILE_TEST support as the driver doesn't compile on
> > non-ARM platforms due to usage of the ARM DMA IOMMU API.
> 
> For similar reasons as x86, can we please think about using:
> 
> 	depends on ARM
> 	depends on ARCH_SHMOBILE || ARCH_SHMOBILE_MULTI || COMPILE_TEST
> 
> This way we don't end up polluting the configuration for non-shmobile
> platforms.  Same goes for other ARM stuff... the number of options is
> getting rather large and we need to think about keeping that in check
> where its easily possible to do so.

I've been asked to move from depends ARM to no dependency at all, and have 
thus settled for ARM || COMPILE_TEST in most cases, and just ARM where the 
driver couldn't be compiled on other platforms.

I'm fine with your proposed option. As I don't want to respin the series 
dozens of time let's first agree on the course of action, I will then repost 
the patches. Mark, you've pushed towards as few platform dependencies as 
possible, what's your opinion on this ?

-- 
Regards,

Laurent Pinchart

Re: [PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 989 bytes --]

On Wed, Oct 30, 2013 at 12:40:12PM +0100, Laurent Pinchart wrote:

> > For similar reasons as x86, can we please think about using:

> > 	depends on ARM
> > 	depends on ARCH_SHMOBILE || ARCH_SHMOBILE_MULTI || COMPILE_TEST

> I'm fine with your proposed option. As I don't want to respin the series 
> dozens of time let's first agree on the course of action, I will then repost 
> the patches. Mark, you've pushed towards as few platform dependencies as 
> possible, what's your opinion on this ?

In general I think we should have whatever the real depedencies are or
COMPILE_TEST (to the extent that they will actually build cleanly on
other targets).  That way only people who explicitly go looking to
compile test things for build coverage (eg, when doing global cleanups
or API updates) need to be bothered by the extra compile test options.

I'd read the above as saying the code needs ARM to build at all and that
the hardware will only ever appear on SHMOBILE.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[Simon Horman]

On Wed, Oct 30, 2013 at 09:28:54AM -0700, Mark Brown wrote:
> On Wed, Oct 30, 2013 at 12:40:12PM +0100, Laurent Pinchart wrote:
> 
> > > For similar reasons as x86, can we please think about using:
> 
> > > 	depends on ARM
> > > 	depends on ARCH_SHMOBILE || ARCH_SHMOBILE_MULTI || COMPILE_TEST
> 
> > I'm fine with your proposed option. As I don't want to respin the series 
> > dozens of time let's first agree on the course of action, I will then repost 
> > the patches. Mark, you've pushed towards as few platform dependencies as 
> > possible, what's your opinion on this ?
> 
> In general I think we should have whatever the real depedencies are or
> COMPILE_TEST (to the extent that they will actually build cleanly on
> other targets).  That way only people who explicitly go looking to
> compile test things for build coverage (eg, when doing global cleanups
> or API updates) need to be bothered by the extra compile test options.
> 
> I'd read the above as saying the code needs ARM to build at all and that
> the hardware will only ever appear on SHMOBILE.

I am curious to know the value of "depends ARM".
Is it to aid the reading that you spelt out?

Re: [PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 742 bytes --]

On Thu, Oct 31, 2013 at 03:03:05PM +0900, Simon Horman wrote:
> On Wed, Oct 30, 2013 at 09:28:54AM -0700, Mark Brown wrote:
> > On Wed, Oct 30, 2013 at 12:40:12PM +0100, Laurent Pinchart wrote:

> > > > For similar reasons as x86, can we please think about using:
> > 
> > > > 	depends on ARM
> > > > 	depends on ARCH_SHMOBILE || ARCH_SHMOBILE_MULTI || COMPILE_TEST

> > I'd read the above as saying the code needs ARM to build at all and that
> > the hardware will only ever appear on SHMOBILE.

> I am curious to know the value of "depends ARM".
> Is it to aid the reading that you spelt out?

That's a function of the compile time dependencies, if ARM is required
to build then it needs to be an unconditional dependency.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[Simon Horman]

On Thu, Oct 31, 2013 at 10:49:45AM -0700, Mark Brown wrote:
> On Thu, Oct 31, 2013 at 03:03:05PM +0900, Simon Horman wrote:
> > On Wed, Oct 30, 2013 at 09:28:54AM -0700, Mark Brown wrote:
> > > On Wed, Oct 30, 2013 at 12:40:12PM +0100, Laurent Pinchart wrote:
> 
> > > > > For similar reasons as x86, can we please think about using:
> > > 
> > > > > 	depends on ARM
> > > > > 	depends on ARCH_SHMOBILE || ARCH_SHMOBILE_MULTI || COMPILE_TEST
> 
> > > I'd read the above as saying the code needs ARM to build at all and that
> > > the hardware will only ever appear on SHMOBILE.
> 
> > I am curious to know the value of "depends ARM".
> > Is it to aid the reading that you spelt out?
> 
> That's a function of the compile time dependencies, if ARM is required
> to build then it needs to be an unconditional dependency.

Thanks, sorry for missing the obvious.

Re: [PATCH v3] iommu: shmobile: Enable the driver on all ARM platforms

[Laurent Pinchart]

[-- Attachment #1: Type: text/plain, Size: 1331 bytes --]

On Wednesday 30 October 2013 09:28:54 Mark Brown wrote:
> On Wed, Oct 30, 2013 at 12:40:12PM +0100, Laurent Pinchart wrote:
> > > For similar reasons as x86, can we please think about using:
> > > 	depends on ARM
> > > 	depends on ARCH_SHMOBILE || ARCH_SHMOBILE_MULTI || COMPILE_TEST
> > 
> > I'm fine with your proposed option. As I don't want to respin the series
> > dozens of time let's first agree on the course of action, I will then
> > repost the patches. Mark, you've pushed towards as few platform
> > dependencies as possible, what's your opinion on this ?
> 
> In general I think we should have whatever the real depedencies are or
> COMPILE_TEST (to the extent that they will actually build cleanly on
> other targets).  That way only people who explicitly go looking to
> compile test things for build coverage (eg, when doing global cleanups
> or API updates) need to be bothered by the extra compile test options.
> 
> I'd read the above as saying the code needs ARM to build at all and that
> the hardware will only ever appear on SHMOBILE.

That's my reading as well, and it makes sense to express it that way to me. 
I'll resubmit the whole series (minus the patches that have been applied 
already, I'll fix those in follow-up patches) with the dependencies expressed 
that way.

-- 
Regards,

Laurent Pinchart

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 490 bytes --]